Semantics data processing

Abstract

Techniques for deriving additional features from input data are described herein. Input data from a plurality of source files are received. One or more features corresponding to the input data, which includes information about semantic types, is identified. The input data is then processed to generate additional features for the input data. New data corresponding to the additional features are then generated and access to the new data is subsequently provided.

What is claimed is: 
     
         1 . A computer-implemented method, comprising:
 processing input data to identify a feature in the input data, the feature corresponding to a subset of the input data and having a semantic type;   obtaining semantic metadata for the feature, the semantic metadata indicating a first semantic context for the feature;   processing the input data with the obtained semantic metadata to:
 identify, in the subset of the input data and based at least in part on a parameter associated with the input data, a first plurality of elements; and 
 aggregate the first plurality of elements by generating, in a manner determined at least in part on the first semantic context, a second element derived from a subset of the first plurality of elements and having a different second semantic context, the subset of the first plurality of elements selected based at least in part on the parameter; and 
   providing, with the parameter, the second element as associated with the parameter.   
     
     
         2 . The computer-implemented method of  claim 1 , wherein the parameter comprises a second feature in the input data, the second feature having a second semantic type. 
     
     
         3 . The computer-implemented method of  claim 1 , wherein:
 the input data includes the parameter; and   the parameter identifies the manner and the subset of the input data to aggregate.   
     
     
         4 . The computer-implemented method of  claim 1 , wherein the manner is determined based at least in part on the first semantic context and a second semantic context corresponding to a different feature in the input data. 
     
     
         5 . A non-transitory computer-readable storage medium having stored thereon executable instructions that, as a result of being executed by one or more processors of a computer system, cause the computer system to at least:
 process input data to identify a subset of the input data, the subset of the input data corresponding to a feature in the input data, the feature having a first semantic type;   obtain metadata for the feature, the metadata being associated with a first semantic context for the feature;   process the input data to determine, based at least in part on the first semantic context, a second feature corresponding to a second semantic context;   generate, from the input data, new data to correspond to the second feature; and   provide access to the new data as associated with a corresponding subset of the input data.   
     
     
         6 . The non-transitory computer-readable storage medium of  claim 5 , wherein the instructions, if executed, that process the input data, further cause the computer system to heuristically determine the metadata based at least in part on the identified subset of the input data. 
     
     
         7 . The non-transitory computer-readable storage medium of  claim 5 , wherein the instructions, if executed, that process the input data, further cause the computer system to determine the second feature based on information other than the first semantic context. 
     
     
         8 . The non-transitory computer-readable storage medium of  claim 5 , wherein the metadata identifies the first semantic context. 
     
     
         9 . The non-transitory computer-readable storage medium of  claim 5 , wherein the instructions, if executed, that generates the new data, further cause the computer system to generate, based at least in part on other metadata associated with the feature, an identifier for the second feature. 
     
     
         10 . The non-transitory computer-readable storage medium of  claim 5 , wherein the instructions, if executed, that provide access to the new data, further cause the computer system to cause processing of the new data by a machine learning algorithm. 
     
     
         11 . The non-transitory computer-readable storage medium of  claim 5 , wherein the instructions, if executed, that provide access to the new data, further cause the computer system to cause processing, by a different computer system, of the new data. 
     
     
         12 . The non-transitory computer-readable storage medium of  claim 5 , wherein the instructions, if executed, that generates the new data, further cause the computer system to determine the second feature using an algorithm identified in a policy as applicable to the feature. 
     
     
         13 . A system, comprising:
 one or more processors; and   memory that stores computer-executable instructions that, if executed, cause the one or more processors to:
 process input data to identify a subset of the input data, the subset of the input data corresponding to a feature in the input data, the feature having a first semantic type; 
 obtain metadata for the feature, the metadata being associated with a first semantic context for the feature; 
 process the input data to determine, based at least in part on the first semantic context, a second feature corresponding to a second semantic context; 
 generate, from the input data, new data to correspond to the second feature; and 
 provide access to the new data as associated with a corresponding subset of the input data. 
   
     
     
         14 . The system  claim 13 , wherein the instructions, if executed, that process the input data, further cause the system to heuristically determine the metadata based at least in part on the identified subset of the input data. 
     
     
         15 . The system of  claim 13 , wherein the instructions, if executed, that process the input data, further cause the system to determine the second feature based on information other than the first semantic context. 
     
     
         16 . The system of  claim 13 , wherein the metadata identifies the first semantic context. 
     
     
         17 . The system of  claim 13 , wherein the instructions, if executed, that generates the new data, further cause the system to generate, based at least in part on other metadata associated with the feature, an identifier for the second feature. 
     
     
         18 . The system of  claim 13 , wherein the instructions, if executed, that provide access to the new data, further cause the system to cause processing of the new data by a machine learning algorithm. 
     
     
         19 . The system  claim 13 , wherein the instructions, if executed, that provide access to the new data, further cause the system to cause processing, by a different computer system, of the new data. 
     
     
         20 . The system of  claim 13 , wherein the instructions, if executed, that generates the new data, further cause the system to determine the second feature using an algorithm identified in a policy as applicable to the feature.